[R] auto.arima and intervention analysis

Gregoire gregory.eady at gmail.com
Mon Mar 5 08:34:42 CET 2012


I'm currently using auto.arima to verify the order of my arima model. I
would like to use the model to conduct an intervention analysis. The problem
is that, when I include a step function in auto.arima, by including a binary
variable in "xreg", the arima order that auto.arima gives is different from
when I don't include it. From my understanding, the "pre-whitening" that is
required for intervention analysis simply means to determine the (p,d,q)
BEFORE adding exogenous regressors. Would auto.arima not do this first, and
only then apply the exogenous regressors? If this is the case, I assumed
that the arima order would be the same for whether I included the binary or
not. This doesn't seem to be the case, so maybe my understanding of
"pre-whitening" is incorrect. Any help would be appreciated.

Thanks so much

View this message in context: http://r.789695.n4.nabble.com/auto-arima-and-intervention-analysis-tp4445324p4445324.html
Sent from the R help mailing list archive at Nabble.com.

More information about the R-help mailing list